Bio::Parser::FileParser    v1.18

  use Bio::Parser::FileParser;
  use vars qw( @ISA );
  @ISA = qw( Bio::Parser::FileParser );

new()

BioParser users should never need to invoke this method but BioParser developers are required to invoke it via SUPER::new() in any new FileParser modules that they create. It in turn invokes the new() constructor from the base Bio::Parser module.

object_mode()

  $fileparser->object_mode(1);

The value of object_mode() changes the behaviour of the next_record() method. A call to next_record() normally returns a text scalar containing the current record but if object_mode() is set to true, then the text record is parsed and a Bio::Parser::XXX::Record object is returned. This method takes a single value that evaluates to true or false - by convention 0, and 1 are used. Setting object_mode() to false causes next_record() to revert to its default behaviour and return the record as a text scalar.

parser()

Store a reference to the RDParser.pm module being used to parse records into objects. A single instance of this module is used as a factory to parse all records.

record_count()

Returns a scalar integer corresponding to the number of records returned so far for this file.

line_count()

Returns a scalar integer corresponding to the number of lines read so far for this file.

record_text()

Returns the current record as a single text scalar. No parsing is done.

record_object()

This method uses the class of the invoking FileParser to guess the classes of the matching RDParser and Record classes. If the FileParser class does not already contain an RDParser instance, one is invoked and used to parse the text returned by a call to record_text(). The hashref returned by RDParser is then blessed into the appropriate Record class and returned. If you are not using an RDParser to do the text-to-object conversion then you'll probably have to provide code to override this method in your FileParser module.

next_record()

This method must be implemented by all Bio::Parser::FileParser subclasses. When called, this method should read a full record from the text file, calling the _incr_line_count() and _incr_record_count() methods as required. It should place the record text into the FileParser object with a call to record_text(). If it is possible to get the current record ID from the text object, then this parameter should also be set with a call to this_id(). For example implementations, see Bio::Parser::OMIM::FileParser and Bio::Parser::LocusLink::FileParser.

next_id()

this_id()

These two methods are optional but useful. They usually have to be implemented by doing a regex on the text record since the parsed record is not necessarily available. These methods should only be implemented if it's easy to extract the IDs of the current and next records without doing the recursive descent parse.

text_file_os()

  my $fh = IO::File->new( 'data_file.txt', 'r' );
  my $os = text_file_os( $fh );

This routines takes as input an open filehandle and returns a string representing the operating system of the file. The string returned is one of 'win', 'unix', 'mac', or 'unknown'.

Fair warning - the routine is implemented using the perl seek, read and tell functions. It stores the current position of the filehandle using tell and reads bytes from the filehandle until it has seen 3 line endings of the same type (win/unix/mac) or until the end of the file is reached. This only works with text files - you'll probably get an answer with a binary file but it will be meaningless. At the end of the reads, the routine use seek to reset the position of the filehandle to whatever it was when the routine was called.

text_file_line_separator()

  my $fh = IO::File->new( 'data_file.txt', 'r' );
  {
      local $/ = text_file_line_separator( $fh );
      my $ctr = 0;
      while (my $line = $fh->getline) {
          chomp $line;
          print $ctr++," [$line]\n";
      }
  }
  $fh->close;

This routine is the same as text_file_os() except instead of returning a string identifying the operating system, it returns the line separator that was found. This can be used as shown in the example above to set a local copy of $/.

_incr_line_count()

Should be called in subclass next_record() methods to keep track of how many lines have been read from the text file.

_incr_record_count()

Should be called in subclass next_record() methods to keep track of how many records have been read from the text file.

_record()

Returns the current record by checking the status of object_mode() and calling either record_object() or record_text() as required.

John Pearson, bioinfresearch@tgen.org